*****************************************************************************************
README File
Supply-Side Innovations to Increase Equitable Access to Digital Financial Services: Experimental Evidence from Mozambique
Mahesh Karra, Mindy Hernandez, Catherine Brennan, Margaret McConnell
*****************************************************************************************

INTRODUCTION
This replication package reproduces the results for the paper entitled "Supply-Side Innovations to Increase Equitable Access to Digital Financial Services: Experimental Evidence from Mozambique," which is published in the Journal of Development Studies (JDS) at XXX.

FILE AND FOLDER SET-UP:
1. Save all data files and the do-file in a folder named DATA.
2. Create a folder named RESULTS in the same directory where the DATA folder is created (but not in the DATA folder).

To run the do-file, open the REPLICATION_DO_FILE.do file in Stata, change the username based on your computer's username (line 12 in the do-file), the filepath for maindir to the directory where the data files are saved (line 13 in the do-file). Execute the do-file in Stata by clicking on Run. The outputs of each file will be stored in the "RESULTS" folder.

1. REPLICATION_DO_FILE.do: The do file that replicates Tables 1-4 in the main paper, Figure 2 in the main paper, Appendix Tables A1, A3-A7, and A9 in the Appendix, Figures B1 and B2 in the Supplemental Materials, and Tables B3 and B4 in the Supplemental Materials. Figure 1, Appendix Tables A2 and A8, and Supplemental Material Tables B1 and B2 do not use the analytic datasets and are produced externally.

		Expected run time: 45 seconds

*****************************************************************************************
The tables are produced using Stata version 14.2 and the following additional commands are used:
- estout: To export outputs from the do-file to figures and tables.
- balancetable: To run a balance table comparing treatment (Female TSR) and control (Male TSR) characteristics.
- itsa: To run the interrupted time series analysis.
- ritest: To run the randomization inference robustness check.
- rwolf2: To run the Romano-Wolf standard error correction for multiple hypothesis testing.

*****************************************************************************************
QUESTIONS?
Please direct any questions regarding this replication package to the corresponding author. 

Correspoding Author:
Mahesh Karra
Heilbrunn Department of Population and Family Health
Columbia University
mvk2109@cumc.columbia.edu